| ||
Wed 2024-11-20
Tue 2024-11-19
Mon 2024-11-18
Sun 2024-11-17
Love, Acceptance, and Screeching Modems Sat 2024-11-16
Fri 2024-11-15
Search
Archives
2023
12 11 10 09 08 07 06 05 04 03 02 01
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
One good site
MDN
Nelson Minar
Blog licensed under a Creative Commons License
|
I have the arrogance of having worked at Google; I thought I knew all there
was about Internet technology leadership. But I've overlooked Facebook;
they're doing some very impressive tech. And Facebook's openness gives
them some mindshare over Google.
Pingdom just posted a roundup of Facebook technologies that's worth reading. Load balancers, distributed datastores, scaleable logging, etc: a lot of good systems technology. Many of their systems tools are available open source; a lot of startups I talk to are using various parts of Facebook's tech. Google has plenty of open source too but they tend to keep the server systems proprietary. If you want a distributed datastore today you may read the BigTable paper but you're going to be using Facebook's Cassandra. Facebook's API strategy is also impressive. They own the American social network: if you're building something using social relationships it's really tempting to do it inside Facebook. (Or as one startup friend of mine put it, "Facebook is layer 8 of the Internet".) People have complained about how awkward the earlier Facebook APIs have been to use. But the Facebook Graph API is the new hotness and it's nicely designed. I'm particularly struck by how it's typeless, all data objects in Facebook have a unique 64 bit id and they're all accessed via URLs like https://graph.facebook.com/649745863. I'm liking this flat REST+JSON design. |
|
